Release 1.3.0
Agent compatibility: In order to benefit from the latest features and improvements, it is necessary to upgrade the App Insights agent to the latest version (1.6). See upgrade instructions further down.
What's new?
Observability & Troubleshooting
Quickly identify problematic workloads! You can quickly identify workloads that are experiencing issues by filtering workloads based on status. Then dive in and see more details for that workload including any insights that can help identify the root cause.
View workload details without direct cluster access - The workload details show a quick summary of replicas & availability, containers & images, resource requests & limits as well as the Kubernetes manifest in YAML form. This allows users to view workload details even if they don’t have direct access to the cluster
Connect the dots & troubleshoot faster - The workload timeline helps visualize recent changes such as image or config changes, cluster issues such as node crashes, memory pressure, auto scaling ..etc and other workloads experiencing issues. This information is valuable when troubleshooting a workload crash or availability issue.
Noisy Neighbours - The Timeline charts now provide a breakdown of pod error counts, enabling you to quickly identify other problematic workloads on your cluster that may adversely affect the workload you are investigating.
Timeline with Bird's-Eye-View & Zoom-In - The timeline provides a 48-hour visualization of events to quickly identify patterns and hot spots. The new “zoom-in” feature allows you to zoom in with 1-minute granularity to dive deeper into a one-hour window.
Get notified of new insights with our Slack, Email and Generic Webhook integrations. This feature is in tech-preview.
Cost Awareness & Optimizations
We firmly believe cost awareness is the key first step to better management of costs. In continuing with that theme,
- Cost Increase Alerts - We compare 24-hour and 7-day periods to identify an increase (or decrease) in cost. We also provide which resources contributed to the cost and the top 10 workloads driving the cost increase.
Other Improvements
- Support for App Insights agent installation & billing through Azure Marketplace
- Breadcrumbs for easier navigation
- Insights List is now sorted by the time the insight was last updated, instead of being sorted by the time the insight was generated
- Cluster metrics and basic info has been moved to a new “Info” tab in the Cluster Details page
See it in action
You can quickly identify workloads that are experiencing issues. For example, the filters above find all workloads that are experiencing a Crash loop back off or that have a new deployment pending due to a failure. You can then check the insights for that workload to identify the underlying issues
Workload overview provides a summary, which includes workload status, replicas & availability, containers & images, resource requests and limits, and the kubernetes manifest for the workload as a yaml.
Receive notifications of new insights with our Slack, Email and Generic Webhook integrations (tech preview)
Whenever an availability issue is identified, you’ll get an insight including recent changes potentially related to the issue, such as image or config changes as well related issues & insights within the last 7 days.
The timeline chart provides a breakdown of pods that have experienced issues allowing you to identify the workloads that may have experienced issue and there by affecting the current workload under investigation.
The Cost Increase insight now displays a workload breakdown with the top 10 workloads that are driving the cost increase
How to update the App Insights agent from version 1.5.* to version 1.6
Updating the App Insights agent from version 1.5.* to version 1.6 can be done with the following steps:
Step 1) Make sure you have the latest version of our Helm repo with the following command:
helm repo update randoli
Step 2) Upgrade the Helm release
Make sure kubectl is configured to use the cluster for which you want to update the App Insights agent. If you want to update multiple clusters, you'll need to repeat this step for each cluster.
Once kubectl is configured, upgrade the Helm release with the following command:
helm upgrade --reuse-values --set image.tag=1.6 insights randoli/app-insights-agent -n randoli-agents
The above command assumes you are using the default installation commands provided by Randoli. If you changed the release name or namespace in your first installation, then adjust the above command accordingly.
Step 3) Verify the agent image
To make sure the correct agent version is deployed, you can check the image:
kubectl get pods -n randoli-agents -o jsonpath='{range .items[*]}{"\n"}{.metadata.name}{":\t"}{range .spec.containers[*]}{.image}{end}{end}'
The expected output would be docker.io/randoli/insights-agent:1.6
Step 4) Verify the Helm release version
To make sure the correct Helm chart is deployed, run the following command and make sure that the APP VERSION is 1.6:
helm list -n randoli-agents
That's it! If you face any issues, please reach out to our support team, who is always happy to help!
How to update the App Insights agent from version 1.4.* to version 1.6
You can check the image tag being used on your App Insights agent deployment. If the image is 1.4, it is necessary to uninstall the previous agent version on all your clusters connected to App Insights, and then install the newest agent, as follows:
Step 1) Uninstall the current agent you have installed:
helm uninstall insights -n randoli-agents
Step 2) Update helm repositories
helm repo update randoli
Step 3) Install the agent again
This last step varies depending on your cluster settings, please login to the App Insights console, go to the Clusters page, select your cluster, press “Install Agent” button, adjust your setting in the 3rd step (Open Cost and Prometheus options), and copy the last command. Here is an example of the helm command using the default options:
helm install insights randoli/app-insights-agent -n randoli-agents --set global.prometheus.url=http://my-prometheus-server.prometheus.svc.cluster.local:80
Step 4) Verify the agent image
To make sure the correct agent version is deployed, you can check the image:
kubectl get pods -n randoli-agents -o jsonpath='{range .items[*]}{"\n"}{.metadata.name}{":\t"}{range .spec.containers[*]}{.image}{end}{end}'
The expected output would be docker.io/randoli/insights-agent:1.6
If you face any issues, please reach out to our support team, who is always happy to help!